High-z Galaxy Formation in LCDM
Theory Challenges
Avishai Dekel, HU Jerusalem,
Santa Cruz, August 2010
LCDM makes certain solid theoretical predictions
for the most active phase of galaxy formation:
massive galaxies (~1011MȨ) at high z (~2-3)
Theory seems consistent with observations,
introducing a coherent picture
Open questions: SFR & feedback
Cosmic Web, Cold Streams, Clumpy Disks & Spheroids
100 kpc
50 Mpc
100 kpc
stars
5 kpc
5 kpc
Bimodality of Stream-Fed
Galaxies
cold
streams
spheroid
thick disk
hot
halo
clumpy
disk
Mv>1012
z>2
clumpy streams
(mergers) build a
spheroid & generate
turbulence →
disk stabilization &
quenching of SFR
smooth
streams
clumpy disk
spheroid
late disk
The main mode of galaxy formation
hot
halo
Theory Challenges
- Galaxies from the cosmic web
- Cold streams in Lyman alpha
- Outflows
- Stream-disk interaction
- Wild clumpy disks
- angular momentum
- driving the turbulence
- clump support
- clump disruption
- Bulge-less disks
- Fate of wild disks - stabilization
Galaxies emerge from the Cosmic Web
- Halos M>>MPS - high-sigma peaks
….at the nodes of the cosmic web
- Typically fed by 3 big streams
- Streams are co-planar
the millenium cosmological simulation
Three Streams: filament mergers
AMR RAMSES
Teyssier, Dekel
box 300 kpc
res 30 pc
z = 5.0 to 2.5
Accretion Rate into a Halo
Neistein, van den Bosch, Dekel 06; Neistein & Dekel 07, 08; Genel et al 08
From N-body simulations/EPS in LCDM (<10% accuracy):
−1
1.14
2 .4
&
M baryon = 80 M ⊗ yr M 12 (1 + z ) 3 f 0.17
Almost all penetrate to the inner halo in cold streams
The accretion rate governs galaxy growth & SFR
- can serve for successful simple modeling
Virial Shock Heating
Birnboim & Dekel 03, Keres et al 05, Dekel & Birnboim 06
−1
cool
t
−1
compress
<t
- Critical halo mass ~1012Mʘ
- Cold streams penetrate at z>2
Kravtsov
At High z, in Massive Halos:
Cold Streams in Hot Halos
in M>Mshock
Totally hot
at z<1
shock
Cold streams
at z>2
cooling
Dekel &
Birnboim 2006
Kravtsov et al
no shock
Cold Streams in Big Galaxies at High z
1014
1013
all hot
Mvir
[Mʘ]
cold filaments
in hot medium
Mshock~M*
1012
Mshock>>M*
Mshock
all cold
1011
M*
1010
0
109
1
2
3
redshift z
4
Dekel &
5 Birnboim 06
high-sigma halos: fed by relatively thin, dense filaments
→ cold narrow streams
typical halos: reside in relatively thick filaments, fed ~spherically
→ no cold streams
the millenium cosmological simulation
Narrow dense gas streams at high z
versus spherical infall at low z
high z
M=1012Mʘ >>MPS
Ocvirk, Pichon, Teyssier 08
low z
M=1012Mʘ ~MPS
Cold Sterams
Neistein et al
06, 07, 08;
Genel et al 08
−1
1.14
2.4
&
M baryon = 80 M ⊗ yr M12 (1 + z )3 f 0.17
- >90% of influx in cold streams
- Penetration: V~Vvir dM/dt(r)~const
- Hot accretion negligible
– recycled outflows
Talks by Keres, Stewart, Leitner
100 kps
Dekel et al 09
Nature
Inflow rate through the halo
into the disk
Cosmological hydro simulations MareNostrum, Dekel et al. 09
Distribution of gas inflow rate
P ( M& | M )
The flow rate is constant with radius:
deep penetration
smooth
flows
mergers >1:10
Galaxy density at a given gas inflow rate
∞
n( M& ) = ∫ P( M& | M ) n( M ) dM
0
P(Mdot|M) from
cosmological hydro
simulations
(MareNostrum)
n(M) by Sheth-Tormen
SFR ~(1/2) inflow rate
M& ≈ 2M& *
Star-Forming Gal’s
Sub-Millimeter Gal’s
Dekel et al 09, Nature
SFR=
Lyman-alpha from Cold streams
Fardal et al 01; Furlanetto et al 05; Dijkstra & Loeb 09
Goerdt, Dekel, Sternberg, Ceverino, Teyssier, Primack 10
T=(1-5)x104 K
n=0.01-0.1 cm-3 NHI~1020 cm-2
L ~1043-44 erg s-1
- Extended cold gas is provided by the incoming streams
- Lya is powered by gravitational infall into halo potential well
100 kpc
Surface brightness
Cold streams as Lyman-alpha Blobs
100 kpc
Goerdt,
Dekel,
Sternberg,
Ceverino,
Teyssier,
Primack 10
Matsuda et al 06-09
Lyman Alpha
Emission (LABs)
Radiative transfer, ionization from stars, dust
Kasen, Fumagalli, Ceverino, Dekel, Prochaska, Primack
- Lya emission from gravitationally infalling
…streams is inavitable in high-z massive halos
- Also from ouflows and ionization by stars or AGN
100 kpc
Massive Outflows Observed
Steidel et al. 2010
Stacked line profiles of Lya and metal absorption
No inflow?
M& out ≈ M& sf → M& in ≈ M& sf + M& out ≈ 2M& out
Weak inflow absorption features in the average line profile
→ small sky coverage → narrow cold streams
blue metal
absorption
red Lya
emission
Talks by Koo, Prochaska
Neutral Hydrogen
Column Density
Radiative transfer, ionization from stars, dust
Kasen, Fumagalli, Ceverino, Dekel, Prochaska, Primack
CIE
NH
NHI
UV bkgd
UV bkgd+stars
Talk by Fumagalli
100 kpc
NHI
NHI
Lya
Absor
HI column density
ption
background source
9%
central source
500
average line profile
Absorption line profile is weak
because of low sky coverage
Inflow signal consistent with
observatios (Steidel et al. 10)
Inflow undetectable in metals
because of low Z and coverage
Inflow and
Outflow
- What drives the massive outflows in massive galaxies?
– Are the cold streams unaffected?
Stream Clumpiness - Mergers
Dekel et al 09, Nature
Mass input to galaxies (all along streams)
- Major mergers >1:3
- Major+minor mergers >1:10
<10%
~33%
- Miniminors and smooth flows ~67%
TalkM=10
by 12
Teyssier
Mʘ z=2.5
Angular Momentum
kpc in the angular momentum
- Streams100
bring
- Disk spin & size are determined by one stream
- Clumpy streams generate turbulence
Talk by Bouche
Open issues:
- Origin of extra-large disk sizes ?
- Origin of “dispersion-dominated” galaxies V/σ<2 ?
..….Angular momentum? Stream clumpiness?
…...Feedback? …Stage of evolution?
Agertz et al 09
A Disk Fed by Cold Streams
streams
disk
interaction
region
Stream-disk interaction? Stream collisions? Stream
instability – hydrodynamical? thermal? gravitational?
Violent Disk Instability
High gas density → disk wildly unstable Q ≈
Giant clumps and transient features
Noguchi 99
Immeli et al. 04
σΩ
≤1
π GΣ
R clump ≈
7G Σ
Ω2
Bournaud,
Elmegreen,
Elmegreen 06, 08
Dekel, Sari,
Ceverino 09
Ceverino, Dekel,
Bournaud 09
Agertz et al. 09
5 kpc
Self-regulation at Q ~ 1 with high σ/V~1/4
Star formation and feedback in the clumps
Talks by
Cacciato,
Ceverino,
Teyssier,
Genel
Rapid migration of massive clumps and mass inflow → bulge formation
What Drives the
Turbulence?
- Gravity: accretion, migration
GM &
&
E≈
M
R
- Feedback: SN, radiative, AGN
Talks by Genel, Cacciato
Clump Support: The Clumps are
Spinning
Jeans equation for an isotropic rotator
Rotation support, induced by disk rotation
and AM conservation during clump collapse,
but the spin can be tilted
2
Vcirc
=
2
rot
2
circ
V
V
GM
= Vrot2 + 2σ 2
R
 Σ clump
≈ 0.2 
 Σ disk
1/ 2



Talk by Ceveriono
Rotating Clumps in a Wildly
Unstable Disk
Naab
Clump Disruption by Stellar Radiation Pressure
Murray et al. 10; Krumholz & Dekel 10
SFR efficiency
Σ& ∗
ε≡
Σ g / tff
~ 0.01 -- Kennicutt law
tff ≈ 15 Myr M 9−1/ 2 R13 / 2
If tff > 3 Myr, the mass fraction ejected is
f eject ≈ 0.08 ε − 2 (Σ −1M 9 ) −1/ 4
Giant clumps in high-z disks survive
if the SFR obeys the Kennicutt law
If ε~0.1, “fireworks”: clumpy disk in steady state:
the clumps are disrupted before they migrate to
the center as new clumps form
A Fireworks Model of
Clumpy Disks
Clumpy disk in steady state: the clumps are disrupted
before they migrate to the center as new clumps form
&
SFR efficiencies η = M c∗ ≈ 0.1 ε = M c∗
M c t ff
Mc
Simulations:
M
α=∑
c
Mdisk
≈ 0.2
M&
∑
β=
M& ∗
c∗
≈ 0.5
η =ε
Simulation steady state:
t form ≈ t mig ≈ 16t ff
Duty cycle
t dis
t dis
µ=
≈ 0.06
t mig
t ff
t dis
t ff
β M& ∗ t dis
η=
α M disk
ε = 0.1 β α 0−.12 ( M& ∗ M d ) Gyr t ff , 20 ≈ 0.1
−1
t dis
t dis ≈ t ff
but should be ~1
Wildly Unstable Disk: Migration
into a Bulge
Noguchi 99; Immeli et al. 04; Bournaud, Elmegreen, Elmegreen 06, 08
Clump Formation & Migration to a Bulge
Cosmological Steady State
Dekel, Sari, Ceverino 09
smooth
streams
migration
γ = 0 - 0.33 - 0.6
M& evac
(1− γ ) M& acc
stream
clumps
γ M& acc
mergers
M& disk = (1− γ) M& acc − M& evac(δ )
M disk
δ≡
M tot
M&
= γ M& + M& (δ )
bulge
acc
evac
Steady state for a few Gyr:
draining disk is replenished by cold streams,
bulge ~ disk ~ dark matter Talks by Cacciato, Genel
Wild Disk Instability Bulge
gas
young stars
dark matter
stars
Observations vs. Simulations
Elmegreen et al
Talk by
Mozena
Stellar Images of Clumpy
Disks z=2
B
H
A typical star-forming galaxy at z=2:
clumpy, rotating, extended disk & a bulge
Hα star-form
regions
color-code
velocity field
Genzel et al 08
Clumpy Disks with Massive Bulges?
A bulge-less disk ???
BzK 15504
z=2.4
BX 482
z=2.2
-200 will soon form a bulge?
- A young clumpy disk that
1”
- Bulge removed by feedback (SN, radiative, AGN)?
(8 kpc)
-240
?
+190
+210
Clump
coalescence
Hα
+ R-band
NIC2
intoHαthe
vels bulge = wet mergers.
If SFR efficiency in mergers
is 10xKennicutt then
BzK-ZC782941 z=2.2
-130
BzK 6004 z=2.4
stellar radiative feedback could disrupt the bulge
+160
-160
+ 140
Genzel et al. 08; Förster Schreiber et al. 10
M(≤3 kpc)/M(≤15 kpc)~0.2-0.4
Hα
rotation
dispersion
Kinema
tics
of
Simul
ated
Clump
y
Disk
Sub-structure in the disk giant clumps
Caution: high-res → substructure → less dissipation
→ smaller collapse factor → less spin-up
Caution: MW molecular clouds are not spin-supported
Bournaud, Teyssier 10 AMR 2 pc resolution
Stabilization by
stream clumps
n ~ 0.01
Dekel, Sari, Ceverino 09
γ = 0 - 0.33 - 0.6
Q≈
dense
stream
clumps
γ M& acc
M tot σ
M disk V
shock
lowdensity
stream
n ~ 0.1
dense
disk
V
- Stabilization Q>1 due to bulge growth & turbulence
…driven by clumpy streams
- Stable disk in steady state for Mdisk/Mtot<0.3
…
→ Bimodality at high z: blue disks and red spheroids
12. Disk Stabilization – SF Quenching
- Dominant bulge – Morpholopgical quenching
- Excessive turbulence by external sources:
…clumpy streams, feedback
- Low accretion rate (e.g. at late times)
Martig et al 09
M tot σ
Q≈
M disk V
- Low gas fraction (e.g. today’s spirals)
Relation to today’s galaxies ?
- The descendants of the high-z clumpy disks are probably
…S0s and rotating Es, or thick disks of spirals
- Thin disks form later by slow accretion
Conclusion
LCDM makes certain solid theoretical predictions for
how massive galaxies form at high z, consistent with
observations, together suggesting a coherent picture
- Galaxies are fed by cold streams from the cosmic web
…Streams include major & minor mergers and smooth flows
…Streams radiate as Lyman-alpha blobs
- Gas-rich disks form, develop violent instability, self-regulated
…Giant clumps form stars (?) and migrate to a bulge. Disruption (?)
…Cosmological steady state with bulge ~ disk. Bulge disruption (?)
…Angular momentum versus dispersion (?)
- Spheroids form by mergers and by violent disk instability
- Disks are stabilized (SFR quenched) by bulge, external turbulence,
…low accretion rate, gas consumption & stellar dominance
- Main open issues: star formation & feedback
Cosmic Web, Cold Streams, Clumpy Disks & Spheroids
100 kpc
50 Mpc
100 kpc
stars
5 kpc
5 kpc